ChatGPT cheats: AI is a bigger threat to educational equality

opinion

Schools and colleges are panicking about artificial intelligence (AI) and cheating. But AI poses a far greater threat to educational equity.

Fear of cheating usually stems from concerns about fairness. How is it fair that one student spends weeks writing an essay while another asks ChatGPT to write the same thing in just a few minutes? It’s essential to worry about giving each student a ‘fair go’ in order to keep the idea going.

But like the ‘American Dream’ myth, New Zealand’s egalitarian narrative masks more pernicious inequalities such as systemic racism and the housing crisis. Both of these have a tremendous and decidedly unfair impact on today’s students.

These persistent inequalities reduce the threat of AI-induced fraud. Educators can benefit from being prepared for other unfairness in AI instead of caring too much about cheating. All of this is demonstrated in OpenAI’s latest large-scale language model (LLM), GPT-4.

GPT-4 Comes at a Price

With sophisticated guardrails and more parameters than ChatGPT, GPT-4 is touted as more secure and accurate than its predecessor. But there are pitfalls. GPT-4 costs $20 USD per month.

For some people, that price doesn’t matter. But for those whose budgets have been squeezed by rapid inflation, it could be a deal breaker. The potential for democratizing AI technology is here, but only if you can afford it.

This information divide divides students and institutions into two camps. People with sufficient resources to reap the benefits of AI tools. And those who don’t have the same financial flexibility to be left behind.

It may seem small right now, but as the cost of AI tools rises, this digital divide could widen into a gigantic chasm. This should worry educators who have long been concerned about how unequal access to learning technology creates inequality among students.

AI Threatens Indigenous Languages ​​and Data

AI tools are also perpetuating the global dominance of English at the expense of other languages, especially oral and indigenous languages. I recently spoke with a Microsoft executive who called these other languages ​​”edge cases.” This is a term used to describe rare cases that cause problems in computer code.

However, indigenous languages ​​are only a “problem” for AI tools, as large language models learn from online data sets that have little indigenous content and predominantly English content.

The dominance of English content online is no accident. English dominates the internet because centuries of British colonization and American cultural imperialism have made English the lingua franca of global capitalism, education and internet discourse. From this point of view, no other language is inferior to English. They don’t make as much money as English content.

But Māori speakers are rightly wary of attempts to commercialize their language. Too often, the commercialization of indigenous knowledge does not benefit indigenous peoples. It is therefore essential that indigenous communities remain in control of their information. This is a concept known as indigenous data sovereignty.

Without indigenous data sovereignty, these multi-billion dollar tech companies could derive value from these so-called edge cases and later decide to stop investing in them.

These threats are important to educators, as AI tools will soon be built into Microsoft Office, search engines, and other learning platforms.

At Massey University, where I teach, students can submit assignments in Maori or English. However, if AI writing tools are better at English than Māori, Māori learners will be at a disadvantage. It also matters if Māori students are forced to use tools that violate the data sovereignty of indigenous peoples.

Banning AI in Education Creates Unfairness

Banning AI in education, as some schools and journals and even some countries are already doing, is tempting, but it also widens existing inequalities. People with disabilities can benefit from communicating with AI tools. But like laptop bans in earlier eras, AI bans prevent students with disabilities from accessing important learning technologies.

Banning AI also penalizes multilingual students who may have trouble writing in English. AI tools help multilingual students learn important English genres, structures, prose styles, grammar, and all the skills that contribute to social mobility. But banning AI puts these multilingual students at a disadvantage.

Instead of banning AI, educators would be better off changing curricula, teaching methods, and assessments to align with soon-to-be widespread AI tools. However, such revisions require more time and resources, and school teachers and college educators have recently taken note of this. Institutions must be prepared to invest not only in AI tools, but also in the educators who are essential to helping students think critically about using AI tools.